Lix builders have IFD disabled, and the changes needed to support this
are minimal.
We do need to do slightly more manual work to keep rust-toolchain.toml
in sync with flake.nix, but it's not significant.
This renames:
database_backend -> database.backend
database_path -> database.path
db_cache_capacity_mb -> database.cache_capacity_mb
rocksdb_max_open_files -> database.rocksdb_max_open_files
Charles updated the NixOS module.
Co-authored-by: Charles Hall <charles@computer.surgery>
This adds some new tools in CI for lint the book and also sets up
automated deployment to GitLab Pages. Also adds a readme that suggests
reading the book, since that's where all the information will be.
We don't support this anyway. OCI images will be needed for testing with
Complement, but there are a bunch of other special requirements that has
so these oci image outputs won't be very useful for that anyway.
This is a config compatibility break.
The ability to listen on multiple ports, including both TLS and non-TLS,
is necessary for running complement against grapevine.
This functionality was never actually used AFAICT, as no way to provide
alternate profiles was ever provided.
This changes the configuration format to remove the `[global]` section,
everything that was previously under that namespace is now at the top
level.
The configuration file is now the canonical way to, well, configure.
This change is desirable because it gives us much more flexibility with
how configuration is structured. Environment variables are insufficient
because, for example, they're a flat namespace and have no built-in way
to represent lists.
Because the nix rocksdb build depends on the jemalloc feature, you need
to use a different devshell when passing --all-features to cargo than
the default.
Some of the features affect nix dependencies, so we need to have a
full feature list available when constructing the nix derivation. This
incidentally fixes the bug where we weren't enabling jemalloc on rocksdb
in CI/devshells, because jemalloc is now a default feature. It does not
fix the more general class of that issue, where CI is performing an
`--all-features` build in a nix devshell built for default-features.
I am now passing `--no-default-features` to cargo, and having it use our
unified feature list rather than duplicating the unification inside cargo.
Without setting JEMALLOC_OVERRIDE, we end up linking to two different
jemalloc builds. Once dynamically, as a transitive dependency through
rocksdb, and a second time to the static jemalloc that tikv-jemalloc-sys
builds.
This fixes dynamically-linked jemalloc builds, for the reasons described
in <https://github.com/girlbossceo/conduwuit/pull/400#issue-2316700200>.
This makes it possible to deploy Grapevine while using a database
originally created by Conduit, including leaving the admin bot user's
localpart the same as before.
This prevents us from needing to recompile the dependencies when that
environment variable changes, which generally changes on every commit,
which is far more frequently than the dependencies are actually changed.
Users of the nix package can now just use `.override` to choose what
features they want.
This also makes RocksDB automatically use jemalloc when Grapevine is
configured to use jemalloc.
